« Culture dominant minorities? | Gene Expression Front Page | Conservatives for Dean? »
January 18, 2004

Language & genes

Abiola notes on his blog that the Fulani speak a Niger-Congo language while the Hausa speak an Afro-Asiatic tongue. When Cavalli-Sforza came out with The History and Geography of Human Genes he noted the striking correlation between genetic markers and languages. This was a throw-back position, before World War II people would use languages and genetic affinity as if they were interchangeable, but this practice went into decline.

Here are the languages associated with Fulani:

The languages of this family include the west African languages of Fulani (Nigeria, Cameroon, Mali, Guinea, Gambia, Senegal, Mauritania, Niger, Burkina Faso), Malinke (Senegal, Gambia, Guinea, Mali, Ivory Coast), Mende (Sierra Leone), Twi (Ghana), Ewe (Ghana, Togo), Mossi (Burkina Faso), Yoruba (Nigeria), Ibo (Nigeria), Kpelle (Liberia), Wolof (Senegal, Gambia) and Fang (Cameroon, Gabon, Guinea).

The languages associated with Hausa include the Berber and Semitic languages (Arabic, etc.).

The pecularity of this is of course that the Hausa are less "Caucasoid" than the Fulani-so it seems that the two groups speak languages they shouldn't, according to the broad Afro-Asiastic = Caucasoid and Niger-Congo = Sub-Saharan African rules (the Fulani being more classically Caucasoid). A quick PubMed search yielded this article which indicates that the Fulani are not very Caucasoid at all according to some genetic markers! That certainly might clear up the original question, but these debates remain in other contexts.

Cavalli-Sforza took data from all around the world, and as a general tendency the concordance between language and genes exists (especially if you correct for the post-1500 spread of European languages). Nevertheless, many people tend to focus on specific people that don't fit into the general predictions, and expect an extremely high level of specificity from the trends. Frankly, they are looking at it deterministically and want to resolve extra-scientific disputes, which often are difficult to adhere to if you couch conclusions probabilistically.

Language groups often closely match national identities, and once you mix nationalism, you get all sorts of strained arguments and debates. Kemal Ataturk, a man that was likely of Slavic and Albanian ancestry, funded research to discern the uniqueness of the Turkish "race." Decades after his death, it is becoming quite clear that the penetration of "Turkish genes" into the Anatolian population was rather low. Now that the idea that Turkey is a "European nation" has solidified among the elite of that country, this data might be something to be proud of. The reality of who the Turkish people are has remained constant, while scientific investigation has become more precise, but the political ramifications and emotional expectations are always a churning sea.

The beauty of an avant guarde field like historical population genetics is that conflicting & ambiguous data, that needs to be set in context, can satisfy everyone! Just pick an outlier study, and ignore everything else-I've seen this many times with ideologues.

Posted by razib at 05:51 PM